Non-repetitive peak forward current

Thread Starter

tiptonian

Joined Jun 30, 2010
2
Non-repetetive peak forward current for a diode is quoted on datasheets for a square pulse of particular current and time. How should you treat other waveshapes, such as a reverse sawtooth (fast rise, slow decay) such as you get through a blocking diode to charge a capacitor via a resistor from a source with a fast rise time?
 

t_n_k

Joined Mar 6, 2009
5,455
One option would be to use the equivalent value of

\(It=\int ^{t}_{0}{i(t)dt}\)

or maybe ???

\(I^2t=\int ^{t}_{0}{i(t)^2dt}\)

for a different waveform.

The latter lends some weighting to the let-through energy concept such as one associates with the rating of semiconductor fuses.
 

The Electrician

Joined Oct 9, 2007
2,849
Non-repetetive peak forward current for a diode is quoted on datasheets for a square pulse of particular current and time. How should you treat other waveshapes, such as a reverse sawtooth (fast rise, slow decay) such as you get through a blocking diode to charge a capacitor via a resistor from a source with a fast rise time?
The most accurate method is to use the transient thermal impedance model for the diode. See:

http://www.pwrx.com/pwrx/app/zth_underchip_nf.pdf

http://www.fairchildsemi.com/an/AN/AN-7522.pdf

and many other references found by searching for the phrase "transient thermal impedance".

Most rectifier diode spec sheets I've looked at specify the non-repetitive surge with a half-sine waveform.

See the attachments, which also discuss derating for low duty cycle.
 

Attachments

Thread Starter

tiptonian

Joined Jun 30, 2010
2
Thanks for the suggestions; unfortunately the device (BAS16L, NXP) does not have any transient thermal impedance data. The SPICE approach is interesting if I could find some values to plug into the circuit, perhaps relating to a mechanically similar part. I shall think on that.
 

nomurphy

Joined Aug 8, 2005
567
Non-repetetive [sic] peak forward current....
Note the term "non-repetitive".

In other words, it is the one-time peak current. I would not try to rationalize it as area under the curve for some other waveform type, when used in a practical circuit, but stay well below the spec.
 
Last edited:

t_n_k

Joined Mar 6, 2009
5,455
For the single non-repetitive current pulse case, a search of the relevant literature might favor the equivalent "integral I squared t" approach rather than transient thermal impedance method.
 
Top